Preoperative planning should include the use of templates of the hardware to be used, superimposed on preoperative radiographs (Fig. 10.2). The normal side may be used as a template for the prospective end result of the surgery. Special implants or prostheses that may be used should be ordered well in advance of the planned surgery. Special tests such as a bone scan, arteriogram, or Doppler ultrasound for vascularity should be performed as part of the preoperative plan.

Consultations with anesthesia, vascular surgery, plastic surgery, internal medicine, cardiology, neurology, urology, psychiatry, and any other services necessary for diagnosis or treatment should be obtained well in advance of surgery.

Overall, preoperative planning should be thorough, thoughtful, and complete. It should include social services with consideration of financial and family concerns, in addition to the medical and surgical consultations listed previously. Only with proper preoperative planning can a successful result be obtained in a high percentage of elderly patients who have a nonunion.

The reported incidence of nonunion following fracture of the femoral neck varies from less than 5% to greater than one third of observed cases within the literature [24]. Its clinical significance is highlighted by the high rates of reoperation for cases of nonunion. In two randomized controlled trials, the rates of nonunion were 34.5% and 36%, with avascular necrosis occurring in 4.9% and 6%, respectively [25, 26]. In these trials, the reoperation rate after internal fixation was 40% and 42%, respectively, with 75% of those progressing to nonunion requiring surgery [25, 26].

In comparison, avascular necrosis of the femoral head following femoral neck fracture requires reoperation in one third of patients [24]. Certain factors have been repeatedly shown to be associated with nonunion of a femoral neck fracture. To date, these factors have yet to reliably demonstrate robust predictive capacity. The association between osteonecrosis of the femoral head and femoral neck fracture nonunion is important to discuss.

Development of osteonecrosis does not preclude the ability to heal a femoral neck nonunion. Prior studies have shown union in the face of established osteonecrosis of the femoral head as well as osteonecrosis that develops following surgical treatment of the nonunion [27, 28].

Certainly, the amount of initial fracture displacement has some bearing on the development of nonunion. There is a clear relationship with fracture nonunion and displaced femoral neck fractures compared to non-displaced fractures [29]. Fracture comminution , particularly of the posterior cortex, has been linked to decrease fixation stability and increased incidence of nonunion [30]. Timing of surgery and adequacy of the reduction may be more important for the development of osteonecrosis but may also contribute to the development of nonunion [27]. Furthermore, the type of implant utilized may have some bearing on the incidence of nonunion [31, 32].

The orthopedic literature is replete with studies that look at treatment of femoral neck fractures. In a systematic review of the literature and a meta-analysis of 106 studies, the incidence of femoral neck nonunion in displaced femoral neck fractures was reported to be as high as 33% [33]. Factors felt to be associated with development of fracture nonunion included inadequate reduction, poor internal fixation, premature weightbearing, and infection [34]. Individual studies of femoral neck fractures observed varied results. Banks reported on the 20-year experience in Boston. He reported on 301 femoral neck fractures, of which 296 were displaced. There were 34 nonunions in this series [35]. In their study of over 1500 femoral neck fractures, Barnes et al. found no correlation with length of time to reduction and the development of nonunion, but the adequacy of reduction was an important factor with regard to development of nonunion [36]. In a prospective, randomized study comparing sliding hip screw and divergent cannulated pins for the treatment of displaced femoral neck fractures, the authors reported an overall incidence of nonunion of 28%, with the sliding hip screw group significantly higher. Interestingly, no correlation between adequacy of reduction and development of nonunion could be made [31].

Yang, Lin, and Chao et al. reported a series of 202 femoral neck fractures treated with three cannulated screws . Their nonunion rate was 21.7% with significant differences in nonunion rate among fracture type, reduction quality, and screw tip subchondral purchase [24]. Their study also reported a significant difference in rate of fracture nonunion based upon cannulated screw configuration. In a similar series reported by Cobb and Gibson, comprised of 65 femoral neck fractures with adequate reduction and technically sound fracture repair, the nonunion incidence was only 4.7% [37]. Garden reported the incidence of nonunion to be 16.6% in his series of 500 patients with subcapital femoral neck fractures [38]. Another report on 76 patients out of an initial 179 showed an incidence of nonunion with and without osteonecrosis of 21% [39].

In a prospective, randomized study looking at internal fixation of femoral neck fractures, 128 patients were treated with a sliding compression screw and 127 with a nail plate. Eleven percent of Garden 3 and 4 fractures went on to nonunion with the compression screw and 25% with the nail plate device [40]. These authors confirmed results of previous studies in which the presence of a varus malunion was associated with development of nonunion. Stromqvist et al. had 22 healing complications in 68 displaced femoral neck fractures. The rate was higher in the flanged nail group [41]. In a second report, Stromqvist et al. reported on 300 femoral neck fractures treated with “Hook Pins.” Of these 215 fractures were displaced, Garden 3 or 4. The incidence of nonunion was 25% overall and 35% in surviving patients [41]. Finally, Skinner and Pauwels reported that in a series of 107 displaced femoral neck fractures, 15 developed a nonunion by 1 year in fractures treated with a sliding hip screw [42].

The incidence of nonunion following surgical repair of an intertrochanteric hip fracture is rare, given the excellent blood supply and cancellous bone stock within the segment. Literature regarding intertrochanteric fracture nonunion and its treatment is limited. Limited case series regarding revision internal fixation and bone grafting following fixation failure have shown encouraging results [35, 43]. Most intertrochanteric fractures treated by conservative methods or internal fixation heal. In a large prospective study performed on over 500 intertrochanteric hip fractures, Kyle et al. reported a 2% incidence of nonunion. All of these occurred in the unstable type 4 fracture patterns treated with a variety of implants [44]. However, nonunion rates may approach 10% when excessive stripping of comminuted fractures disturbs bone nutrition [45]. Mariani and Rand reported on the treatment of ununited intertrochanteric hip fractures. Upon reviewing the initial postoperative radiographs, the authors noted an association with failed ORIF and poor reduction as well as medial displacement osteotomy [46]. Bogoch et al. reported a 6.5% incidence of nonunion following intertrochanteric hip fracture in patients with rheumatoid arthritis [47]. The authors felt the condition of rheumatoid bone was a contributing factor in healing complications.

Subtrochanteric nonunion is more common than intertrochanteric nonunion, most likely owing to its high-stress region in the femur. The incidence of nonunion reported in the literature range from 0.5 to 5% [48–50]. Nonunion of subtrochanteric fractures may be related to poor fracture reduction, unfavorable fracture pattern, poor bone quality, loss of medial column support, and early weightbearing [51–53].

Seinsheimer reported the results of 56 patients treated for subtrochanteric fracture treated with a variety of methods and implants. There were eight failures of fixation and three persistent nonunions reported [49]. The authors associated failures and nonunions with fractures that had extensive comminution in the posteromedial cortex; thus, the lateral plate is subjected to excessive medial bending forces as well as a fracture length greater than 8 cm.

Zickel reported on 84 subtrochanteric femur fractures treated by one surgeon with a cephalomedullary device . He had only one nonunion in his series [54]. Wiss et al. reported on 95 subtrochanteric femur fractures. There was only one nonunion in his series (1%) of fractures all treated with an interlocking intramedullary nail [50]. The authors performed their surgery without exposing the fracture site, and thus there was no stripping of medial soft tissues.

In their series of 50 patients with subtrochanteric femur fractures, Velasco and Comfort reported a complication rate of 21%. However, only one patient (2%) developed nonunion of their fracture [55].